Clickitongue is a tool that removes the need to physically click mouse buttons, to help people without full use of their fingers, or severe pain from repetitive-strain-injured wrists. Its goal is for an experienced mouse user to feel no more encumbered than they would normally clicking with their fingers. Clickitongue learns to recognize specific sounds, and issues clicks (left and right) in response. It can tolerate noisy environments - people talking, air conditioners running nearby - without issuing spurious clicks.
Clickitongue is available to everyone free of charge under the GPLv3. It runs on Linux (both X and Wayland), Windows, and OSX.
On the first run, Clickitongue will have you select your audio input device, and then guide you through its training process. The training is quick; it should take about two to five minutes total.
Just about any microphone, including webcam or built-in laptop mics, should get you basic functionality. However, for really smooth operation the mic needs to be positioned near your mouth. More details here.